Analyzing recurrent event data with informative censoring

Recurrent event data are frequently encountered in longitudinal follow-up s tudies. In statistical literature, noninformative censoring is typically as sumed when statistical methods and theory are developed for analyzing recur rent event data. In many applications, however, the observation Of recurren t events could be terminated by informative dropouts or failure events, and it is unrealistic to assume that the censoring mechanism is independent of the recurrent event process. In this article we consider recurrent events of the same type and allow the censoring mechanism to be possibly informati ve. The occurrence of recurrent events is modeled by a subject-specific non stationary Poisson process via a latent variable. A multiplicative intensit y model is used as the underlying model for nonparametric estimation of the cumulative rate function, The multiplicative intensity model is also exten ded to a regression model by taking the covariate information into account. Statistical methods and theory are developed for estimation of the cumulat ive rate function and regression parameters. As a major feature of this art icle, we treat the distributions of both the censoring and latent variables as nuisance parameters. We avoid modeling and estimating the nuisance para meters by proper procedures. An analysis of the AIDS Link, to Intravenous E xperiences cohort data is presented to illustrate the proposed methods.